摘要
为保护油浸式配电变压器(以下简称“配变”)安全稳定运行,提出基于多模态数据融合的油浸式配变分级预警模型。首先,采用历史负荷数据训练支持向量回归模型,预测配变超短期负荷。其次,分析油浸式配变的温度特性,提出油顶温度预测和绕组热点温度计算的方法,其中绕组热点温度计算中引入了空气流速修正因子来提高计算结果准确性。然后,提出计及过载倍率、过载时长、热点温度的多模态数据融合的配变分级预警模型,在二维平面上确定3个等级的预警线,考虑温度效应建立三维运行区域的配变分级预警划分方法,将配变的运行状态描述为过载倍率、过载时长、绕组热点温度三维坐标的运行点,通过分析运行点在运行区域的位置来确定预警级别。最后,通过实际算例分析验证了所提模型的有效性和可靠性。
In order to protect safe and stable operation of oil-immersed distribution transformers,this paper proposes a graded early warning model based on multi-modal data fusion.Firstly,the support vector regression model is used to train the historical load data to predict the ultra-short term load of the distribution transformer.Secondly,the temperature characteristics of oil-immersed distribution transformers are analyzed,and the methods for oil top temperature prediction and winding hot spot temperature calculation are proposed,among which the air velocity correction factor is introduced into the calculation of the winding hot spot temperature to improve the accuracy of the calculation results.Then,the paper proposes a grading early warning model for the distribution transformer,which takes into account overload ratio,operating time and hot spot temperature.Three levels of early warning lines are determined on the two-dimensional plane and then the temperature effect is considered to establish a three-dimensional operating area classification method for early warning of the distribution transformer.The operating state of the distribution transformer is described as the operating point of the three-dimensional coordinates of overload ratio,operating time,and winding hot spot temperature,and the warning level is determined by analyzing the position of the operating point in the operating area.Finally,the paper verifies the validity and reliability of the proposed model through example analysis.
作者
石墨
许海林
李晖
罗颖婷
江俊飞
SHI Mo;XU Hailin;LI Hui;LUO Yingting;JIANG Junfei(Electric Power Research Institute of Guangdong Power Grid Co.,Ltd.,Gaungzhou,Guangdong 510080,China)
出处
《广东电力》
2022年第12期110-117,共8页
Guangdong Electric Power
基金
广东电网有限责任公司科技项目﹝GDKJXM20199485(036100KK52190066)。
关键词
油浸式配电变压器
支持向量回归
油顶温度
绕组热点温度
预警
oil-immersed distribution transformers
support vector regression
oil top temperature
winding hot spot temperature
early warning